基于APACHE II评分，Lactate和Fibrinogen 联合构建脓毒症患者预后预测模型

Construction of a Prognostic Prediction Model for Patients with Sepsis Based on the Combination of APACHE II Score, Lactate and Fibrinogen

摘要: 目的：探讨影响重症监护室(ICU)脓毒症患者28天预后的独立危险因素，并构建一个稳定、可靠的预测模型。方法：采用病例对照研究设计，回顾性收集2017~2022年中国人民解放军联勤保障部队第九四〇医院ICU收治的脓毒症患者的临床资料，根据28天生存状态分为生存组与死亡组。经正态性检验和方差齐性检验确定统计方法，通过组间差异分析筛选出P < 0.05 的变量。建立二元多因素Logistic回归模型并进行列线图可视化，采用受试者工作特征(ROC)曲线评估模型的区分度。进一步运用DeLong检验、净重新分类指数(NRI)和综合判别改善指数(IDI)评估联合模型相较于单一指标的预测效能改善情况，并通过决策曲线分析(DCA)评价模型的临床净获益。结果：共纳入290例脓毒症患者，其中生存206例(71.0%)、死亡84例(29.0%)。单因素分析显示，高血压史(HTN)、糖尿病史(DM)、年龄、急性生理学与慢性健康状况评分系统II (APACHE II)评分、乳酸(lactate)、纤维蛋白原(fibrinogen)与预后显著相关(P < 0.05)。多因素Logistic回归显示，APACHE II评分lactate及fibrinogen是脓毒症死亡的独立影响因素。联合指标构建的ROC曲线下面积(AUC)为0.785，DeLong检验、NRI和IDI均显示联合模型的预测效能显著优于各单一指标。DCA结果显示，模型在0.08~0.91阈值范围内具有临床净获益。结论：基于APACHE II评分、lactate及fibrinogen三个指标构建的脓毒症预后预测模型，具有良好的区分度及临床实用性，可为ICU脓毒症患者早期风险评估提供参考。

Abstract: Objective: To investigate the independent influencing factors for the 28-day prognosis of patients with sepsis in the intensive care unit (ICU) and to develop a stable and reliable predictive model. Methods: A case-control study was conducted. Clinical data of patients with sepsis admitted to the ICU of the 940th Hospital of the Joint Logistics Support Force of the Chinese People’s Liberation Army from 2017 to 2022 were retrospectively collected. Patients were divided into survivor and non-survivor groups according to their 28-day survival status. Statistical methods were determined after normality and homogeneity of variance tests, and variables with a P value < 0.05 were retained following between-group difference analysis. A binary multivariate logistic regression model was constructed and visualized using a nomogram. The discrimination ability of the model was evaluated by the receiver operating characteristic (ROC) curve. DeLong’s test, net reclassification improvement (NRI), and integrated discrimination improvement (IDI) were further applied to assess the improvement in predictive performance of the combined model compared with individual predictors, and decision curve analysis (DCA) was performed to evaluate the clinical net benefit of the model. Results: A total of 290 patients with sepsis were included, of whom 206 survived (71.0%) and 84 died (29.0%). Univariate analysis revealed that hypertension, diabetes mellitus, age, Acute Physiology and Chronic Health Evaluation II (APACHE II) score, lactate, and fibrinogen were significantly associated with prognosis (P < 0.05). Multivariate logistic regression identified APACHE II score, lactate, and fibrinogen as independent influencing factors for sepsis-associated mortality. The area under the ROC curve (AUC) of the combined model was 0.785. DeLong’s test, NRI and IDI all demonstrated that the combined model exhibited significantly better predictive performance than each individual predictor. DCA showed that the model provided a favorable net clinical benefit across a threshold range of 0.08~0.91. Conclusion: The predictive model based on APACHE II score, lactate, and fibrinogen exhibits good discrimination ability and clinical utility, offering a valuable reference for early risk assessment in ICU patients with sepsis.